role of argp (icia) in lysine-mediated repression in ... gowrishankar_publications... · specific...
TRANSCRIPT
Published Ahead of Print 2 September 2011. 2011, 193(21):5985. DOI: 10.1128/JB.05869-11. J. Bacteriol.
Carmelita N. Marbaniang and J. Gowrishankar Repression in Escherichia coliRole of ArgP (IciA) in Lysine-Mediated
http://jb.asm.org/content/193/21/5985Updated information and services can be found at:
These include:
REFERENCEShttp://jb.asm.org/content/193/21/5985#ref-list-1at:
This article cites 46 articles, 32 of which can be accessed free
CONTENT ALERTS more»articles cite this article),
Receive: RSS Feeds, eTOCs, free email alerts (when new
http://jb.asm.org/site/misc/reprints.xhtmlInformation about commercial reprint orders: http://journals.asm.org/site/subscriptions/To subscribe to to another ASM Journal go to:
on Novem
ber 10, 2011 by CE
NT
RE
FO
R D
NA
FIN
GE
RP
RIN
TIN
G A
ND
DIA
GN
OS
TIC
S (C
DF
D)
http://jb.asm.org/
Dow
nloaded from
JOURNAL OF BACTERIOLOGY, Nov. 2011, p. 5985–5996 Vol. 193, No. 210021-9193/11/$12.00 doi:10.1128/JB.05869-11Copyright © 2011, American Society for Microbiology. All Rights Reserved.
Role of ArgP (IciA) in Lysine-Mediated Repression in Escherichia coli�
Carmelita N. Marbaniang and J. Gowrishankar*Laboratory of Bacterial Genetics, Centre for DNA Fingerprinting and Diagnostics, Hyderabad 500 001, India
Received 22 July 2011/Accepted 24 August 2011
Initially identified as an inhibitor of oriC-initiated DNA replication in vitro, the ArgP or IciA protein ofEscherichia coli has subsequently been described as a nucleoid-associated protein and also as a transcriptionalregulator of genes involved in DNA replication (dnaA and nrdA) and amino acid metabolism (argO, dapB, andgdhA [the last in Klebsiella pneumoniae]). ArgP mediates lysine (Lys) repression of argO, dapB, and gdhA in vivo,for which two alternative mechanisms have been identified: at the dapB and gdhA regulatory regions, ArgPbinding is reduced upon the addition of Lys, whereas at argO, RNA polymerase is trapped at the step ofpromoter clearance by Lys-bound ArgP. In this study, we have examined promoter-lac fusions in strains thatwere argP� or �argP or that were carrying dominant argP mutations in order to identify several new genes thatare ArgP-regulated in vivo, including lysP, lysC, lysA, dapD, and asd (in addition to argO, dapB, and gdhA). Allwere repressed upon Lys supplementation, and in vitro studies demonstrated that ArgP binds to the corre-sponding regulatory regions in a Lys-sensitive manner (with the exception of argO, whose binding to ArgP wasLys insensitive). Neither dnaA nor nrdA was ArgP regulated in vivo, although their regulatory regions exhibitedlow-affinity binding to ArgP. Our results suggest that ArgP is a transcriptional regulator for Lys repression ofgenes in E. coli but that it is noncanonical in that it also exhibits low-affinity binding, without apparent directregulatory effect, to a number of additional sites in the genome.
The functional role of the ArgP or IciA protein in Esche-richia coli is an enigma in that it has been variously describedas a canonical transcriptional activator, an inhibitor of chro-mosome replication initiation, and a nucleoid-associated pro-tein. First identified as a protein that binds specific sequencesin oriC in vitro to inhibit the initiation of replication (21–23,46), ArgP was shown (46) to be a member of the family ofLysR-type transcriptional regulators (LTTRs) and, subse-quently, to be essential both in vivo and in vitro for transcrip-tion of the gene encoding the L-arginine (Arg) exporter ArgO(26, 32); in Corynebacterium glutamicum, LysG and LysE areorthologous to E. coli ArgP and ArgO, respectively, and LysGis required for LysE transcription (5). ArgP-regulated tran-scription in vivo has also been demonstrated for the genesdapB of E. coli (7) and gdhA of Klebsiella pneumoniae (15).(That E. coli gdhA may be under ArgP control had also beensuggested earlier [33].) In addition, in vitro studies have sug-gested that ArgP activates the transcription of the dnaA andnrdA genes that are involved in DNA metabolism and replica-tion (20, 27–29).
At the same time, ArgP has also been reported as a nucle-oid-associated protein that shows apparently sequence-non-specific DNA binding activity (2). The protein exhibits affinityfor AT-rich and curved DNA sequences, as determined in vitroby DNase I footprinting or electrophoretic mobility shift assays(EMSAs) (2, 48). ArgP exists as a dimer at an estimated con-centration of around 100 to 400 molecules per cell (that is,around 200 to 800 nM monomers) (3, 23).
As mentioned above, one of the well-characterized targets
for transcriptional regulation by ArgP is the argO gene (26, 32,37). Transcription of argO in vivo is activated both by Arg andits toxic analog L-canavanine, both effects being mediated byArgP. On the other hand, argO transcription is drastically re-duced in medium supplemented with L-lysine (Lys), to a levelequivalent to that observed in �argP mutants; when both Argand Lys are present, it is the Lys effect on argO which predom-inates. Dominant mutations in argP have been identified (des-ignated argPd) that confer an L-canavanine-resistant phenotypeand act in trans to considerably increase argO transcription invivo over that obtained in the argP� strain (32).
In vitro, the ArgP protein exhibits high-affinity binding to theargO regulatory region in the presence of either coeffector, Argor Lys, as well as in their absence (7, 26, 37). Stable recruit-ment by ArgP of RNA polymerase to the argO promoter isobserved in the presence of Arg or Lys, following which thetwo coeffectors exert dramatically opposite effects in the ter-nary complex; whereas productive transcription from the argOpromoter is stimulated in the presence of ArgP and Arg, RNApolymerase is trapped at the promoter at a step after opencomplex formation in the presence of ArgP and Lys (26).
As with argO, in other ArgP-regulated promoters, such asdapB of E. coli and gdhA of K. pneumoniae, Lys supplementa-tion is associated with a decrease in transcription in vivo. Inthese cases, however, the addition of Lys results in a decreasedaffinity of ArgP to bind to the corresponding regulatory regionsin vitro, suggesting that it is the failure of RNA polymeraserecruitment at the promoters which is responsible for Lys re-pression (7, 15).
In this study, we have used the �argP and argPd mutations toexamine ArgP’s role in the regulation in vivo of a number ofgenes in E. coli that were chosen on the basis of either theiridentification in the earlier reports or their roles in Lys and Argmetabolism. EMSA experiments to determine ArgP bindingwith the upstream regulatory regions of these genes in vitro
* Corresponding author. Mailing address: Centre for DNA Finger-printing and Diagnostics, Building 7, Gruhakalpa, 5-4-399/B, Nam-pally, Hyderabad 500 001, India. Phone: 91-40-2474 9445. Fax: 91-40-2474 9448. E-mail: [email protected].
� Published ahead of print on 2 September 2011.
5985
on Novem
ber 10, 2011 by CE
NT
RE
FO
R D
NA
FIN
GE
RP
RIN
TIN
G A
ND
DIA
GN
OS
TIC
S (C
DF
D)
http://jb.asm.org/
Dow
nloaded from
were also undertaken. Our results indicate that negative reg-ulation involving Lys as coeffector is mediated by ArgP andthat the target genes include argO, lysP, lysC, asd, dapB, dapD,lysA, and gdhA. Of the genes listed above, argO was unique inseveral respects, including its responsiveness to Arg, ArgPbinding even in the presence of Lys, and behavior with theargPd mutations. Other genes, such as dnaA and nrdA, didexhibit binding to ArgP in vitro (that was Lys insensitive) butwere not regulated by ArgP in vivo. These findings implicate arole for ArgP in Lys metabolism and homeostasis in E. coli.
MATERIALS AND METHODS
Growth media, bacterial strains, and plasmids. The routine defined and richgrowth media were, respectively, glucose-minimal A (with amino acids supple-mented at 40 �g/ml or otherwise as indicated below) and LB medium, as de-scribed previously (31), and the growth temperature was 37°C. The proportion ofacid and basic phosphate in minimal A medium was suitably adjusted (42) toobtain a growth medium of pH 5.8. Ampicillin (Amp), kanamycin (Kan), andspectinomycin (Sp) were each used at 50 �g/ml, and trimethoprim (Tp) at 30�g/ml. L-Arabinose (Ara) supplementation was at 0.2%.
The E. coli K-12 strains used in the study are listed in Table 1. Plasmidspreviously described include the following (salient features are in parentheses):pBAD18 (pMB9 replicon, Ampr, for Ara-induced expression of target genes)(19), pMU575 (IncW single-copy-number replicon with promoterless lacZ gene,Tpr) (1), pCP20 (pSC101-based Ts replicon encoding Flp recombinase, Ampr)(14), and pHYD1723 (pMU575 derivative with 402-bp argO promoter fragmentupstream of lacZ, Tpr) (26). Plasmids bearing argPd alleles encoding one of theArgP variants A68V, S94L, P108S, V144M, P217L, or R295C and the corre-sponding argP� (pHYD915) and vector (pCL1920) controls (pSC101 replicons,Spr) were as described in Nandineni and Gowrishankar (32).
The promoter-lac fusion derivatives that were constructed using plasmidpMU575 in this study are described in Table 2. Plasmid pHYD2673 is a pBAD18derivative carrying the lysA� gene from E. coli (genomic coordinates 2975628 to2976968 [40]) cloned downstream of Para of the vector; the gene was PCRamplified from genomic DNA with the primer pairs 5�-ACAAGGTACCTTTTATGATGTGGCGT-3� and 5�-ACAATCTAGAAGTCATCATGCAACC (KpnIand XbaI sites, respectively, are italicized in the two primers), and the KpnI-XbaI-digested product was cloned into the corresponding sites of pBAD18. Theconstruction of plasmid pHYD2606 encoding the ArgPd P274S variant is de-scribed below.
Construction of mutant encoding ArgP(P274S). The argPd(P274S) mutant hadpreviously been described by Celis (10) as one exhibiting dominant L-canavanineresistance but by a proposed mechanism different from that involving activationof argO transcription. Since all the other argPd mutations confer L-canavanineresistance by the latter mechanism alone (32), we sought to construct and test theP274S variant of ArgP in this study.
TABLE 1. List of E. coli K-12 strains
Straina Genotypeb
MC4100............�(argF-lac)U169 rpsL150 relA1 araD139 flbB5301 deoC1 ptsF25GJ4748 .............MC4100 argR64 zhb-914::Tn10dCmGJ9602 .............MC4100 �argPGJ9623 .............MC4100 �lysPGJ9624 .............MC4100 �argP �lysPGJ9647 .............MC4100 �cadC::KanGJ9648 .............MC4100 �argP �cadC::KanGJ9649 .............MC4100 �lysP �cadC::KanGJ9650 .............MC4100 araD�
GJ9651 .............MC4100 araD� �argPGJ9652 .............MC4100 araD� �lysR::KanGJ9653 .............MC4100 araD� �argP �lysR::Kan
a Strains described earlier (32) include MC4100 and GJ4748. All other strainswere constructed in this study.
b Genotype designations are as described previously (6). All strains are F�.The �argP, �lysP, �lysR, and �cadC alleles were introduced as Kanr deletion-insertion mutations from the Keio knockout collection (4), and where necessary,the Kanr marker was then excised by site-specific recombination with the aid ofplasmid pCP20, as described previously (14). The latter mutations are shownwithout the Kanr designation in the table.
TA
BL
E2.
Lis
tof
pMU
575-
deri
ved
prom
oter
-lac
fusi
onpl
asm
ids
cons
truc
ted
inth
isst
udya
Plas
mid
Prom
oter
�fra
gmen
tsi
ze(b
p)�
Prim
er1
(enz
yme)
Prim
er2
(enz
yme)
Ext
ent
(gen
omic
coor
dina
tes)
b
pHY
D26
02gd
hA(4
35)
5�-A
TT
TT
GA
TC
CT
GC
AG
AA
CG
CA
GC
AC
TG
-3�
(P)
5�-G
TT
TG
AT
TC
GG
AT
CC
CG
CT
TT
TG
GA
CA
TG
-3�
(B)
�32
2to
�11
3(1
8400
09–1
8404
43)
pHY
D26
10da
pD(3
73)
5�-C
GC
CA
TT
TT
AC
TG
CA
GA
AA
CC
GA
AG
-3�
(P)
5�-C
GG
GT
AA
CG
GA
TC
CT
GC
AT
TG
GC
T-3
�(B
)�
273
to�
100
(185
877–
1862
49)
pHY
D26
36ly
sP(2
38)
5�-G
CG
CT
TT
CT
GC
AG
TA
TT
GC
GA
TC
C-3
�(P
)5�
-TA
GT
TT
CG
GA
TC
CC
AT
AC
AA
AA
AT
GC
-3�
(B)
�20
6to
�32
(224
6550
–224
6787
)pH
YD
2647
lysP
(158
)5�
-AC
AA
CT
GC
AG
TT
CG
CC
AG
AA
AA
-3�
(P)
5�-T
AG
TT
TC
GG
AT
CC
CA
TA
CA
AA
AA
TG
C-3
�(B
)�
114
to�
32(2
2465
38–2
2466
95)
pHY
D26
48ly
sP(1
20)
5�-A
CA
AC
TG
CA
GC
TG
GG
CG
AT
CA
T-3
�(P
)5�
-TA
GT
TT
CG
GA
TC
CC
AT
AC
AA
AA
AT
GC
-3�
(B)
�76
to�
32(2
2465
38–2
2466
57)
pHY
D26
58ar
tJ(4
16)
5�-A
CA
AC
TG
CA
GC
AA
AG
CG
CT
GG
CA
-3�
(P)
5�-A
CA
AG
GA
TC
CG
GA
TA
GG
TG
GC
TG
AA
A-3
�(B
)�
270
to�
146
(899
703–
9001
18)
pHY
D26
59ar
tP(2
88)
5�-A
CA
AC
TG
CA
GG
AA
TC
GC
TA
AC
GC
C-3
�(P
)5�
-AC
AA
GG
AT
CC
TA
TC
GA
AC
AG
CG
CC
-3�
(B)
�19
2to
�96
(902
896–
9031
83)
pHY
D26
60ar
gT(3
67)
5�-A
CA
AC
TG
CA
GA
TC
TC
TT
TG
CC
CG
C-3
�(P
)5�
-AC
AA
GG
AT
CC
GT
AG
CG
CC
GC
AT
A-3
�(B
)�
236
to�
131
(242
5740
–242
6106
)pH
YD
2661
hisJ
(306
)5�
-AC
AA
CT
GC
AG
TG
TC
TA
CG
GT
GA
CT
G-3
�(P
)5�
-AC
AA
GG
AT
CC
TC
GC
AG
CA
AA
CG
-3�
(B)
�18
7to
�11
9(2
4247
40–2
4250
45)
pHY
D26
64ly
sC(6
15)
5�-A
CA
AC
TG
CA
GG
TC
TG
CG
TT
GG
AT
T-3
�(P
)5�
-AC
AA
TC
TA
GA
CG
GT
TC
AT
GG
CG
T-3
�(X
)�
244
to�
371
(423
1191
–423
1805
)pH
YD
2668
asd
(343
)5�
-GT
AT
GT
TT
CA
GT
GT
CG
AC
AT
GA
AA
AT
AG
-3�
(S)
5�-G
GC
GT
CG
AA
GC
TT
CG
CT
CT
TC
AA
CC
A-3
�(H
)�
207
to�
136
(357
2827
–357
3169
)pH
YD
2669
dapB
(386
)5�
-CC
CT
GT
TT
TG
CT
GC
AG
TG
GA
AA
C-3
�(P
)5�
-AA
TG
CC
AG
CG
GA
TC
CT
GA
AT
CA
AC
-3�
(B)
�28
5to
�10
1(2
8057
–284
42)
pHY
D26
70ly
sA(2
68)
5�-A
CA
AC
TG
CA
GC
CT
CA
GT
CA
GG
CT
TC
-3�
(P)
5�-A
CA
AG
GA
TC
CC
AG
CC
AA
AT
TC
AG
C-3
�(B
)�
139
to�
129
(297
6844
–297
7111
)pH
YD
2671
dnaA
(668
)5�
-AC
AA
CT
GC
AG
TT
TC
AT
GG
CG
AT
T-3
�(P
)5�
-AC
AA
GG
AT
CC
GC
TG
GT
AA
CT
CA
T-3
�(B
)�
460
to�
208
(388
1696
–388
2363
)pH
YD
2672
nrdA
(481
)5�
-AC
AA
CT
GC
AG
CG
CC
AT
CA
AC
AA
T-3
�(P
)5�
-AC
AA
GG
AT
CC
GT
CG
AG
AT
TG
AT
GC
-3�
(B)
�31
1to
�17
0(2
3424
65–2
3429
45)
pHY
D26
74ca
dB(5
27)
5�-T
TT
AA
TT
TA
CC
TG
CA
GG
GG
CA
AA
C-3
�(P
)5�
-TG
CA
AT
AC
CG
GA
TC
CC
AT
CA
TA
TT
A-3
�(B
)�
387
to�
140
(435
7989
–435
8515
)pH
YD
2675
lysR
(268
)5�
-AC
AA
CT
GC
AG
CA
GC
CA
AA
TT
CA
GC
-3�
(P)
5�-A
CA
AG
GA
TC
CC
CT
CA
GT
CA
GG
CT
TC
-3�
(B)
�12
7to
�14
1(2
9768
44–2
9771
11)
aPr
ovid
edin
the
tabl
efo
rea
chla
cfu
sion
plas
mid
isin
form
atio
non
prim
ers
used
for
PCR
ampl
ifica
tion
ofth
ein
dica
ted
prom
oter
frag
men
tan
dth
een
zym
es(B
,Bam
HI;
H,H
indI
II;P
,Pst
I;S,
SalI
;and
X,X
baI)
used
for
clon
ing
into
plas
mid
vect
orpM
U57
5(1
);th
eco
rres
pond
ing
rest
rict
ion
site
sin
the
prim
ers
are
italic
ized
.b
The
exte
ntof
each
clon
edin
sert
regi
on,r
elat
ive
toth
est
art
site
oftr
ansc
ript
ion
mos
tpr
oxim
alto
the
codi
ngre
gion
ofth
ege
ne,i
sgi
ven;
geno
mic
coor
dina
tes
for
the
clon
edre
gion
sar
efr
omR
udd
(40)
.
5986 MARBANIANG AND GOWRISHANKAR J. BACTERIOL.
on Novem
ber 10, 2011 by CE
NT
RE
FO
R D
NA
FIN
GE
RP
RIN
TIN
G A
ND
DIA
GN
OS
TIC
S (C
DF
D)
http://jb.asm.org/
Dow
nloaded from
Site-directed mutagenesis of the argP� gene on plasmid pHYD915 toargP(P274S) was done with the aid of the QuikChange kit and associated pro-tocol (Stratagene, La Jolla, CA), using the complementary primer pairs 5�-GCACCGCTTTGCTTCTGAAAGCCGCAT-3� and 5�-ATGCGGCTTTCAGAAGCAAAGCGGTGC-3� (nucleotide substitutions are italicized). The resultantplasmid, designated pHYD2606, conferred dominant L-canavanine resistance inan argO� strain but not in a �argO mutant (data not shown), confirming that thisArgP variant also works through ArgO in exhibiting its phenotype.
Microarray expression profiling. For microarray expression profiling experi-ments, RNA preparations were made from the following pair of strains aftergrowth to mid-exponential phase in glucose-minimal A medium with 1 mM Arg(relevant genotypes are in parentheses): (i) strain MC4100 carrying the plasmidpHYD926 encoding ArgPd(S94L) (argP�/argPd); and (ii) strain GJ9602 (�argP).The microarray data were generated at an outsourced facility (Genotypic Tech-nology Pvt. Ltd., Bengaluru, India).
Other methods. The protocols for EMSA experiments with purified ArgPprotein bearing a C-terminal His6 tag (His6-ArgP) have been described earlier(26), and the reactions were performed in EMSA binding buffer of the followingcomposition: 10 mM Tris-Cl at pH 7.5, 1 mM EDTA, 50 mM NaCl, 5 mMdithiothreitol, and 5% glycerol. The DNA templates were obtained by PCR fromE. coli genomic DNA with the various primer pairs listed in Table 2 and oneadditional control template (295-bp fragment from the lacZ locus, genomiccoordinates 364776 to 365070 [40]) with the primer pair 5�-GTGGTGCAACGGGCGCTGGGTCGGTTAC-3� and 5�-CAACTCGCCGCACATCTGAACTTCAG-3�; after 5�-end labeling, the PCR fragments were purified by electroelu-tion following electrophoresis on 6% polyacrylamide gels (42).
Tolerance or sensitivity to L-canavanine was tested by the method given inNandineni et al. (33), while resistance to thialysine was scored with 200-�g/mlsupplementation in glucose-minimal A medium. Procedures for P1 transduction(31) and for PCR, in vitro DNA manipulations, and transformation (42) were asdescribed previously. �-Galactosidase assays were performed by the method ofMiller (31); each value reported is the average of at least three independentexperiments, and the standard error was 10% of the mean in all cases.
RESULTS
Microarray profiling to identify candidate genes regulatedby ArgP in vivo. We initially undertook a microarray-basedcomparison of whole-genome mRNA abundances between anisogenic pair of strains that was either �argP (GJ9602) orexpressed the dominant constitutively acting S94L variant ofthe ArgP protein in the argP� strain (MC4100/pHYD926),which had been grown to exponential phase in Arg-supple-mented glucose-minimal A medium. At the time that theseexperiments were done, the genes known to be ArgP regulatedin vivo included the following (gene functions are indicatedin parentheses): argO (Arg export), dapB (dihydrodipicoli-nate reductase in the diaminopimelate-Lys biosyntheticpathway), and K. pneumoniae gdhA (glutamate dehydroge-nase for NH4
� assimilation and glutamate biosynthesis). Allthree genes are repressed upon growth in Lys-supplementedmedium (7, 15, 32).
In the microarray expression experiments between the �argPand argPd(S94L) strains, the mRNA abundances of all threegenes listed above exceeded the log2 difference threshold of1.0, with values of 2.3, 2.7, and 1.3, respectively. Other genesimplicated in Lys metabolism (36) that crossed this thresholdwere (gene function and log2-fold difference are in parenthe-ses) lysP (Lys permease, 2.3), lysC (Lys-sensitive aspartokinase,1.9), and asd (aspartate semialdehyde dehydrogenase, 1.1).
Accordingly, the E. coli genes argO, dapB, gdhA, lysP, lysC,and asd were chosen for further analysis of in vivo regulation bythe ArgP protein and its constitutive variants. Two additionalgenes, lysA (diaminopimelate decarboxylase) and dapD (tetra-hydrodipicolinate synthase), were also included for the study,
since both are repressed by Lys, as with the known ArgP-regulated genes (36).
Promoter-lac fusion studies in argP� and argP mutants. Foreach of the genes dapB, gdhA, lysP, lysC, asd, lysA and dapD,the upstream cis regulatory region was PCR amplified andcloned into pMU575, which is a single-copy-number IncWreplicon plasmid vector for generating promoter-lacZ tran-scriptional fusions (1). The equivalent argO-lac fusion deriva-tive of pMU575 (pHYD1723) has been described earlier (26).
The eight lac fusion constructs were then introduced into a�argP strain carrying either argP� or any of the constitutiveargP mutant alleles on the pSC101-based plasmid vectorpCL1920; similar derivatives of �argP with vector pCL1920alone served as the negative controls in these experiments.�-Galactosidase assays were performed for cultures of thestrains grown in medium without or with supplementation withArg or Lys at 1 mM. The data from these experiments, pre-sented in Table 3, permitted the following interpretations.
(i) Transcription from promoters of the genes gdhA, asd,dapB, dapD, and lysP in cultures of the argP� strains grown inmedium not supplemented with Arg or Lys is higher than thatin the corresponding �argP derivatives, by factors of approxi-mately 4, 2.5, 23, 4, and 35, respectively. Arg supplementationwas without effect for any of these genes.
(ii) All the genes listed above also exhibited repression uponLys supplementation in the argP� but not the �argP strain,suggesting that Lys repression of their promoters is ArgP me-diated. (iii) As has also been reported earlier (26, 32, 37),argO-lac expression was stimulated 4-fold in the argP� strainrelative to that in �argP but only in Arg-supplemented me-dium. Lys supplementation was associated with repression onlyin the argP� strain.
(iv) The constitutive argP mutations that were used in thisstudy had been selected to confer L-canavanine resistance (thatis, for enhanced expression of ArgO, which is the exporter ofArg and L-canavanine) (10, 32); as expected, all of them dis-played much higher levels of argO-lac expression than theargP� strain without or with Arg or Lys supplementation. TheP274S variant in particular was the most effective for constitu-tive argO expression (with a 270-fold degree of activation).
(v) In the case of the other ArgP-regulated genes discussedabove, many of the argP constitutive mutations behaved muchlike the argP� allele itself for both activation and Lys repres-sion, although a few combinations exhibited differences; forexample, the levels of activation obtained with the P108S vari-ant at gdhA (7-fold) and lysC (5-fold) were higher than thatwith argP� at these promoters, but P108S was also less effectivefor Lys repression of lysP, and both V144M and P217L wereineffective for activation of asd and dapD. However, the mostconspicuous discrepancy was that seen with the P274S variant;although it was the most proficient of all the ArgPd mutants forconstitutive argO expression, this variant was among the leasteffective for activation of expression from all the other ArgP-regulated promoters and, indeed, behaved like �argP at severalof them (gdhA, asd, dapD, and lysC).
(vi) For lysC, there was a 5-fold repression by Lys in the�argP strain (that is, ArgP independent); this presumably rep-resents the regulation imposed by the Lys-responsive ribo-switch that is proposed to exist in the leader region of the lysCmRNA (39, 44). However, the argP� strain displayed 4-fold
VOL. 193, 2011 ArgP AND LYSINE REPRESSION IN E. COLI 5987
on Novem
ber 10, 2011 by CE
NT
RE
FO
R D
NA
FIN
GE
RP
RIN
TIN
G A
ND
DIA
GN
OS
TIC
S (C
DF
D)
http://jb.asm.org/
Dow
nloaded from
higher expression in the unsupplemented medium than the�argP strain and 10-fold repression by Lys, suggestive of anadditional component of control that was ArgP dependent.The argPd(P274S) mutant behaved like the �argP strain,whereas the other argPd derivatives exhibited the additionalcomponent of activated expression and Lys repression, as wasobtained in the argP� strain.
(vii) Along the same lines as with lysC, the expression of thelysA-lac fusion was repressed 6- to 8-fold in the �argP andargPd(P274S) mutants, which reflects the regulation conferred
by the regulator protein LysR (see below). All other argPd
derivatives, as well as the argP� strain, showed 10- to 16-foldrepression upon Lys supplementation.
EMSA studies with ArgP. To determine whether the in vivoexpression differences between argP� and argP mutant strainsfor the various promoter-lac fusion constructs were associatedwith ArgP binding to the corresponding cis regulatory regionsin vitro, we performed EMSA experiments as described below.Earlier studies have shown that ArgP-mediated Lys repressioncan occur by two alternative mechanisms at different promot-
TABLE 3. Expression of lac fusions in different argP variant derivativesa
ArgPvariant
Expression shown as �-Gal sp act (Miller units)b or indicated ratio
argO argT asd
Nil Arg Lys Arg/Lysc Variant/�d Nil Arg Lys Nil/Lys Variant/� Nil Arg Lys Nil/Lys Variant/�
� 38 31 32 1.0 1.0 208 218 204 1.0 1.0 265 284 302 0.8 1.0� 41 106 25 4.2 3.4 214 194 205 1.0 1.0 608 632 281 2.1 2.2A68V 2,248 2,613 1,036 2.5 84 211 197 211 1.0 1.0 447 443 209 2.1 1.6S94L 3,615 3,227 3,776 0.9 104 227 221 224 1.0 1.0 478 450 265 1.8 1.8P108S 1,621 1,633 755 2.2 53 196 200 194 1.0 0.9 871 760 283 3.0 3.2V144M 737 930 176 5.3 30 184 187 189 0.9 0.8 331 344 229 1.4 1.2P217L 2,126 2,436 623 3.9 79 216 215 211 1.0 1.0 352 283 157 2.2 1.3P274S 7,884 8,246 8,392 1.0 266 187 197 206 0.9 0.8 195 207 186 1.0 0.7R295C 115 256 688 2.7 8.3 191 197 196 0.9 0.9 662 675 295 2.2 2.4
dapB dapD dnaA
� 100 108 140 0.7 1.0 161 152 162 0.9 1.0 88 89 95 0.9 1.0� 2,344 2,380 401 5.8 23 642 572 213 3.0 3.9 85 94 90 0.8 0.8A68V 1,074 1,199 289 3.7 11 437 322 130 3.3 2.7 87 87 82 1.0 0.9S94L 1,561 1,468 299 5.2 16 672 567 249 2.6 4.1 72 71 85 0.8 0.8P108S 2,891 2,882 421 6.8 28 659 757 238 2.7 4.0 78 78 91 0.8 0.8V144M 1,692 1,592 305 5.5 17 229 278 134 1.7 1.4 98 85 85 1.1 1.1P217L 1,143 1,024 275 4.1 11 262 216 120 2.1 1.6 92 90 95 0.9 1.0P274S 190 173 127 1.4 1.9 160 139 144 1.1 0.9 87 85 83 1.0 0.9R295C 3,717 3,071 572 6.4 37 301 230 152 1.9 1.8 93 87 90 1.0 1.0
gdhA lysA lysC
� 211 198 192 1.0 1.0 493 535 80 6.1 1.0 553 521 107 5.1 1.0� 801 711 307 2.6 3.7 1,004 980 77 13 2.0 2,095 1,974 214 10 3.7A68V 743 749 313 2.3 3.5 984 980 71 14 1.9 1,477 1,436 197 7.4 2.6S94L 700 687 314 2.2 3.3 910 1,300 71 13 1.8 1,806 1,719 214 8.4 3.2P108S 1,502 1,401 383 3.9 7.1 1,075 1,038 68 16 2.1 2,928 2,736 250 12 5.2V144M 614 555 278 2.2 2.9 782 836 77 10 1.5 1,264 1,268 158 8.0 2.2P217L 609 612 314 1.9 2.8 841 820 64 13 1.7 1,677 1,610 186 9.0 3.0P274S 253 219 177 1.4 1.1 720 511 92 7.8 1.4 678 646 165 4.1 1.2R295C 1,512 1,360 331 4.5 7.1 893 892 69 13 1.8 2,399 2,282 236 10 4.3
lysP nrdA
� 51 50 48 1.0 1.0 338 348 377 0.8 1.0� 1,731 1,592 325 5.3 34 377 357 400 0.9 1.1A68V 1,558 1,455 587 2.6 31 419 399 391 1.0 1.2S94L 1,561 1,653 473 3.3 30 361 331 409 0.8 1.0P108S 1,378 1,434 922 1.4 27 386 398 426 0.9 1.1V144M 1,174 1,136 251 4.6 23 421 413 429 0.9 1.2P217L 881 858 166 5.3 17 340 418 377 0.9 1.0P274S 499 465 169 2.9 10 374 369 399 0.9 1.1R295C 1,938 1,883 987 1.9 38 389 417 387 1.0 1.1
a Derivatives of the �argP strain GJ9602 each carrying two plasmids, as follows, were used in the experiments. (i) Plasmid vector pCL1920 (�) or its derivatives withargP� (�; pHYD915) or the different argPd variants indicated. (ii) One of the following lac fusion plasmids: pHYD1723 (argO), pHYD2660 (argT), pHYD2668 (asd),pHYD2669 (dapB), pHYD2610 (dapD), pHYD2671 (dnaA), pHYD2602 (gdhA), pHYD2670 (lysA), pHYD2664 (lysC), pHYD2636 (lysP), or pHYD2672 (nrdA).
b Specific activities of �-galactosidase (�-Gal) after growth in glucose-minimal A medium supplemented with 18 amino acids other than Arg or Lys (Nil), 18 aminoacids and 1 mM Arg (Arg), or 18 amino acids and 1 mM Lys (Lys) are reported.
c Data indicate the degree of Lys repression for the lac fusion strain concerned, relative to Arg supplementation in the case of argO and to Nil supplementation forthe rest.
d Data indicate the degree of regulation imposed by the ArgP variant concerned (or the wild-type ArgP) relative to that in the �argP strain, in medium with Argsupplementation in the case of argO and with Nil supplementation for the rest.
5988 MARBANIANG AND GOWRISHANKAR J. BACTERIOL.
on Novem
ber 10, 2011 by CE
NT
RE
FO
R D
NA
FIN
GE
RP
RIN
TIN
G A
ND
DIA
GN
OS
TIC
S (C
DF
D)
http://jb.asm.org/
Dow
nloaded from
ers. At dapB or K. pneumoniae gdhA, ArgP binding to theregulatory region is diminished upon the addition of Lys (7,15). At argO, on the other hand, Lys-liganded ArgP binds theregulatory region (7, 26, 37) but it then inhibits productivetranscription at a step further downstream by trapping RNApolymerase at the promoter (26). Accordingly, in this study,EMSA experiments with His6-ArgP were done in the absenceor presence of 0.1 mM Lys, and the cis regulatory regions ofgdhA, lysC, dapB, dapD, asd, lysA, lysP, and argO were em-ployed. The data from these experiments are shown in Fig. 1.The apparent dissociation constants (Kds) for binding of ArgPto the different templates in the presence or absence of Lys aregiven in Table 4.
Consistent with the data from earlier reports (7, 26, 37),ArgP exhibited high-affinity binding to argO (Kd 15 nM) thatwas not diminished in the presence of Lys. Under identicalexperimental conditions as for argO, the regulatory regions ofthe other ArgP-regulated genes (asd, dapB, dapD, gdhA, lysA,lysC, and lysP) were also bound by ArgP, with apparent Kdsranging from 55 nM to 170 nM; in all these cases (unlike thesituation with argO), the addition of Lys was associated with anincrease in the apparent Kd, indicating that ArgP binding inthese instances is Lys sensitive.
Coregulation of gdhA in vivo by both ArgP and NaCl. Inaddition to its role in NH4
� assimilation, glutamate synthesisthrough the two pathways involving glutamate synthase on theone hand and glutamate dehydrogenase on the other (theseare encoded by gltBD and gdhA, respectively) contribute to E.coli osmoregulation, that is, adaptation to growth at high os-molarity (13, 38). We found that gdhA-lac expression is re-duced 3-fold upon growth in a medium rendered highly osmo-lar by NaCl supplementation; furthermore, the decrease inexpression at high osmolarity was additive to that imposed bya �argP mutation, such that gdhA-lac transcription in the �argPmutant grown with NaCl supplementation was 10-fold lowerthan that in the wild-type strain grown without such supple-mentation (Table 5). Although the mechanism of gdhA tran-scriptional regulation by osmolarity remains to be determined,these data may provide an explanation for the earlier findingsof Nandineni et al. (33) that the gltBD argP double mutants aregrowth inhibited at high osmolarity.
Analysis of lysP regulation by ArgP. The results describedabove indicated that ArgP binds the lysP regulatory region tomediate its nearly 35-fold transcriptional activation and thatthe absence of such ArgP binding either in �argP mutants orupon Lys supplementation in argP� strains results in very lowlevels of lysP expression. In support of this conclusion was ourfinding that the �argP strain is resistant to the Lys analogthialysine to the same extent (200 �g/ml) as a �lysP mutant(data not shown).
Recently, Ruiz et al. (41) had also independently identifiedArgP’s role both as a transcriptional activator and in mediatingthe repression of the lysP gene by Lys. However, they reportedthat ArgP binds the lysP regulatory region with similar affinitiesin both the absence and presence of Lys and had accordinglysuggested that ArgP mediates Lys repression of lysP by thesame mechanism of RNA polymerase trapping that it does atargO. Since our results were different from those of Ruiz et al.(41), we repeated the EMSA experiments with lysP, using agraded series of ArgP concentrations in the absence or pres-
ence of Lys (Fig. 2A); the data clearly establish that binding ofArgP to the lysP template is Lys sensitive. The specificity ofArgP’s binding to labeled lysP was also established in thisexperiment (Fig. 2A), by demonstrating that it could be com-peted by unlabeled lysP DNA (lane 12) but not by two othernonspecific DNA fragments (lanes 13 and 14).
When nested deletion constructs of the lysP regulatory re-gion were tested, a fragment extending upstream up to �114bp (relative to the start site of transcription) was nearly indis-tinguishable from the larger fragment (extending up to �219bp) in terms of both in vivo lac fusion regulation (Table 6) andin vitro Lys-sensitive ArgP binding (Fig. 2, compare panels Aand B). On the other hand, a smaller fragment, extending up to�76 bp, was inactive for lac fusion expression in vivo (Table 6)and lysP binding in vitro (Fig. 2C), suggesting that the lysPupstream region between �114 bp and �76 bp carries criticalsequence determinants for binding of and regulation by ArgP.
ArgP and cadBA regulation. The LysP permease has alsobeen implicated in transcriptional regulation of the cadBAoperon encoding Lys decarboxylase. cadBA expression is in-duced only in medium of low pH that is also supplementedwith Lys and is dependent on an activator, CadC. In lysP-defective mutants, cadBA is induced at low pH even in theabsence of Lys, and the model is that LysP negatively regulatescadBA by sequestering CadC in the absence of Lys (34, 45).Since our findings indicated that ArgP is needed for activationof lysP transcription, we tested whether cadBA expressionwould be rendered Lys independent even in the �argP mutant.
Accordingly, a cadBA-lac fusion was constructed and shownto exhibit regulation by cadC, lysP, low pH, and Lys as reportedearlier (34, 45); however, the �argP mutant failed to pheno-copy the �lysP strain, in that cadBA expression continued to beLys dependent in the former (Table 7). These data suggest thatthe basal level of LysP permease present in a �argP strain issufficient for the negative regulation of cadBA in the absenceof Lys.
Testing for cross-regulation by LysR and ArgP genes. Asmentioned above, ArgP belongs to the LTTR family, of whichLysR is the prototypic member (30). Since both LysR andArgP are involved in mediating the repression by Lys of dif-ferent genes, we tested for the possibility of cross-regulation bythe two transcription factors in vivo.
A �lysR strain fails to express lysA and hence is a Lys auxo-troph (36). Therefore, to test for regulation in the presence orabsence of Lys in the �lysR mutant, we constructed derivativesin which lysA was ectopically expressed from the Para promoteron a plasmid (pHYD2673). The LysR-regulated lysA-lac andArgP-regulated lysP-lac fusions were employed for tests of invivo cross-regulation (by ArgP and LysR, respectively); thedata reported in Table 8 indicate that there is no cross-regu-lation of lysP by LysR and that ArgP’s ability to activate lysAoccurs only in a lysR� background. The expression of a lysR-lacfusion was also unaffected in a �argP strain, although it wasincreased 2-fold in a �lysR mutant (Table 8), indicative ofnegative autoregulation (36).
Genes for Arg import are not regulated by ArgP. Since ArgPis required for transcription of the Arg exporter ArgO, we alsotested the promoters of other genes known to be involved inArg transport (artP, artJ, argT, and the hisJQMP operon) ascandidates for regulation by ArgP. The respective promoter-
VOL. 193, 2011 ArgP AND LYSINE REPRESSION IN E. COLI 5989
on Novem
ber 10, 2011 by CE
NT
RE
FO
R D
NA
FIN
GE
RP
RIN
TIN
G A
ND
DIA
GN
OS
TIC
S (C
DF
D)
http://jb.asm.org/
Dow
nloaded from
FIG. 1. EMSAs with ArgP and cis regulatory regions of different genes in the absence or presence of the coeffector Lys. ArgP monomerconcentrations are indicated for each lane. Bands corresponding to free DNA and to DNA in binary complex with ArgP are marked by filled andopen arrowheads, respectively.
5990 MARBANIANG AND GOWRISHANKAR J. BACTERIOL.
on Novem
ber 10, 2011 by CE
NT
RE
FO
R D
NA
FIN
GE
RP
RIN
TIN
G A
ND
DIA
GN
OS
TIC
S (C
DF
D)
http://jb.asm.org/
Dow
nloaded from
lac fusions were constructed, but neither the �argP mutationnor Lys supplementation affected the expression of any ofthem (Table 9). Consistent with earlier reports (8, 9), the lacfusions with the promoters of hisJ, artJ, and artP exhibitedrepression by Arg in the wild-type strain and derepressed ex-pression in a �argR mutant (Table 9). EMSA experimentsundertaken with the regulatory region of one of the genes,argT, revealed a very low affinity of binding by ArgP that wasLys independent (Fig. 3 and Table 4); the expression of theargT-lac fusion was also not affected in any of the argPd mu-tants (Table 3).
ArgP and regulation of genes of DNA replication and me-tabolism. In its description as IciA, ArgP has been reported toregulate dnaA and nrdA transcription (20, 27–29). We con-structed and tested the promoter-lac fusions for both thesegenes, but neither was altered in expression in the �argP orargPd mutants compared to its expression in the argP� strain,either without or with Lys supplementation (Table 3). Ourconclusion runs contrary to that of Han et al. (20), who over-produced IciA (ArgP) and employed an RNase protectionassay to show that nrdA transcription is ArgP regulated in vivo.In EMSA experiments using ArgP and the cis regulatory re-gions of both dnaA and nrdA, we observed binding at an ap-parent Kd of around 150 nM that was unchanged in the pres-ence or absence of Lys (Fig. 3 and Table 4).
Finally, since our data showed that ArgP binds several DNAtemplates in vitro without a corresponding regulatory effect invivo, we asked whether at high concentrations it would exhibitbinding to a completely nonspecific control DNA fragment.When tested by EMSA, negligible binding of ArgP was ob-served even at 400 nM to an internal 295-bp DNA sequencefrom lacZ (Fig. 3 and Table 4). These findings are consistentwith other reports that ArgP does not bind DNA nonspecifi-cally (23, 41), as also with both the inability of nonspecificDNA fragments to compete with lysP for binding to ArgP (Fig.2) and the very weak binding of ArgP to argT (Fig. 3).
DISCUSSION
The Arg exporter ArgO was among the first whose expres-sion in vivo was identified as being under the transcriptionalcontrol of the LTTR ArgP, and the argPd mutants were ob-tained as derivatives with greatly elevated argO expression(32). In the case of argO, ArgP mediates its transcriptionalactivation by Arg, as well as its repression by Lys (7, 26, 32, 37);the latter is achieved by a mechanism involving the activetrapping at the argO promoter by ArgP of RNA polymerase,which is then prevented from being released to engage inproductive transcription (26). Subsequently, dapB and K. pneu-moniae gdhA have also been shown to be ArgP regulated andto be repressed with Lys; in both these instances, Lys repres-sion is achieved by reduced binding of ArgP to the DNAregulatory regions in the presence of Lys (7, 15).
In the present work, we have constructed lac fusions topromoters of genes connected with Arg or Lys metabolism andtransport and studied them in the argP�, �argP, and argPd
strains to identify several new ArgP-regulated genes in vivo.These include lysP, lysC, lysA, dapD, and asd (in addition toargO, dapB, and gdhA), all of which exhibited ArgP-mediatedrepression upon Lys supplementation in vivo, as well as ArgPbinding to the corresponding cis regulatory regions in vitro.
Insights from the use of argPd mutants. For many of thenewly identified targets of ArgP listed above, the magnitude ofin vivo regulation by ArgP (and of repression by Lys) was onlyaround 3-fold, and it was the data on lac fusion expression inthe argPd mutants that offered additional confidence that thisregulation was indeed real.
Although the argPd mutants had been obtained (10, 32) onthe basis of their resistance to L-canavanine and increasedexpression in them of argO, the mutants exhibited differenteffects on the other target genes with regard both to the degreeof activation relative to that in the �argP strain and to thedegree of repression upon Lys supplementation (Table 3). Ofthese, the most significant difference was that for the P274Svariant, which was almost completely defective for activation oflysC, asd, dapD, gdhA, and dapB and partially so for lysP,whereas it was indeed the most effective of all the argPd mu-tants for argO activation. On the other hand, for the genes suchas argT, dnaA, or nrdA that are not ArgP regulated, there wasno difference in expression between the �argP derivative andany of the argPd mutants.
The expression data with the argPd mutants also indirectlyillustrate the complexity of ArgP’s binding interactions withand at the cis regulatory regions of its target genes, since themutations appear to affect these interactions differently at dif-ferent targets.
TABLE 5. Regulation of gdhA expression by ArgP and NaCl
Strain (genotype)�-Gal sp act (Miller units)a
�NaCl �NaCl
MC4100 (argP�) 1,016 331GJ9602 (�argP) 340 102
a Values reported are the specific activities of �-galactosidase (�-Gal) in de-rivatives of the indicated strains carrying gdhA-lac on plasmid pHYD2602 aftergrowth in glucose-minimal A medium without (�NaCl) or with (�NaCl) sup-plementation with 0.4 M NaCl and 1 mM glycine betaine.
TABLE 4. Apparent Kds of ArgP binding to cis regulatory regionsof different genesa
cis regulatory regionKd (nM)
�Lys �Lys
argO 15 15argT �400 �400asd 170 �200dapB 120 �150dapD 70 120dnaA 150 150gdhA 80 130lysA 150 210lysC 70 110lysP 55 �140nrdA 150 150lacZ �400 �400
a From the EMSA autoradiograph for each of the DNA templates, theamounts of radioactivity in the bands corresponding to free (unbound) DNAwere densitometrically determined for the different lanes, and the apparentKd in the absence (�) or presence (�) of 0.1 mM Lys was calculated from thecurve-plotting values of the unbound fraction versus the ArgP concentrationsused.
VOL. 193, 2011 ArgP AND LYSINE REPRESSION IN E. COLI 5991
on Novem
ber 10, 2011 by CE
NT
RE
FO
R D
NA
FIN
GE
RP
RIN
TIN
G A
ND
DIA
GN
OS
TIC
S (C
DF
D)
http://jb.asm.org/
Dow
nloaded from
Regulation by ArgP of Lys-biosynthetic genes. Of the ArgP-regulated and Lys-repressed genes identified earlier and in thisstudy, five encode enzymes for biosynthesis of diaminopimelicacid and Lys (36): two in the pathway that is common for Lys,threonine, and methionine biosynthesis (lysC and asd) andthree in the Lys-specific pathway (dapB, dapD, and lysA). Ofthese, lysC and lysA exhibit additional mechanisms for Lysrepression, involving, respectively, a postulated riboswitch (39,44) and the LysR regulator protein (30, 36).
Our results indicate that ArgP activates lysA transcription2-fold but only in cells that are also proficient for LysR. On theother hand, ArgP-mediated regulation of lysC is independentof and additive to the putative riboswitch regulatory mecha-nism. One may speculate that Lys-dependent modulation byArgP of multiple genes of the pathway provides a fine-tuningeffect in controlling the flux of intermediates serving the bio-synthesis of three different amino acids for protein synthesis, asalso of diaminopimelic acid for peptidoglycan assembly.
FIG. 2. EMSAs with ArgP and the full-length cis regulatory region of lysP (A) or one of its two deletion derivatives (B, C) in the absence (Nil)or presence (Lys) of the coeffector Lys. The extent in base pairs of the lysP regulatory region used in each experiment is given in parentheses. ArgPmonomer concentrations are indicated for each lane. Bands corresponding to free DNA and to DNA in binary complex with ArgP (the latter isnot observed in panel C) are marked by filled and open arrowheads, respectively. Lanes 12 to 14 of panel A depict the results of the addition ofa 100-fold excess of unlabeled competitor DNA, either specific, that is the full-length lysP fragment (S), or one of two different nonspecific DNAfragments (NS1 and NS2), to the mixtures for EMSA reactions undertaken with 160 nM ArgP. NS1 is a 253-bp fragment from the ilvG locusobtained by PCR with the primers 5�-CAGCAACACTGCGCGCAGCTGCGTGATG-3� and 5�-CTTGTGCGCCAACCGCCGCCGGTAAAC-3�(genomic coordinates 3949570 to 3949822 [40]), while NS2 is the same 295-bp lacZ fragment that was used for the EMSA whose results are shownin Fig. 3.
5992 MARBANIANG AND GOWRISHANKAR J. BACTERIOL.
on Novem
ber 10, 2011 by CE
NT
RE
FO
R D
NA
FIN
GE
RP
RIN
TIN
G A
ND
DIA
GN
OS
TIC
S (C
DF
D)
http://jb.asm.org/
Dow
nloaded from
Role for ArgP in Lys-mediated repression in E. coli. Asmentioned above, all genes so far identified to be ArgP medi-ated are repressed by Lys, and we suggest that all Lys-ligandedrepression is mediated by ArgP in E. coli. LysR also mediatesLys repression (of lysA), but in this case it has been suggestedthat the coeffector ligand is diaminopimelic acid (which is thesubstrate for lysA) and that the latter’s binding to LysR con-verts it into its activator conformation (30, 36). Although Lys-liganded repression is a common feature of all the ArgP-reg-ulated genes, different mechanisms operate to achieve suchrepression at different genes, as discussed below.
Regulation of lysP by ArgP. Of the various ArgP-regulatedgenes, the three that exhibited �10-fold activation were argO,dapB, and lysP. The lysP gene encodes a Lys-specific permease,and our data indicate that it is around 35-fold activated byArgP and 5-fold repressed by Lys; a �argP mutant phenocopiesa �lysP mutant for resistance to the Lys analog thialysine. LysRdoes not participate in lysP transcriptional regulation. In vitro,ArgP binds the lysP regulatory region with a Kd of around 55nM and the binding affinity is diminished upon the addition ofLys (to a Kd of �140 nM), suggesting that Lys represses lysP invivo by engendering the loss of ArgP binding to the lysP oper-ator region. The in vitro data also indicate that the lysP up-stream sequence between �76 bp and �114 bp is required forArgP binding.
In addition to its role in active uptake of Lys, the LysPpermease also participates in Lys-dependent transcriptional
regulation of the cadBA operon, by sequestering the CadCactivator in the absence of Lys but not in its presence (34, 45).We found in this study that the �argP mutant does not phe-nocopy a �lysP mutant for cadBA expression, suggesting thatthe basal level of LysP which is present in a �argP strain issufficient for mediating the sequestration of CadC in cadBAregulation.
While this work was being completed, Ruiz et al. (41) inde-pendently reported the identification of lysP as a Lys-repressedtranscriptional target of ArgP. While most of our results onlysP are similar to those of Ruiz et al. (41), one importantdifference is their description that ArgP’s binding to the lysPregulatory region in vitro is Lys insensitive (and therefore sim-ilar to ArgP’s binding to argO), whereas we found that it is Lyssensitive. Perhaps this discrepancy may be related to differ-ences in the protocols employed in the two laboratories forArgP purification or the subsequent protein-DNA bindingstudies, although we emphasize that our demonstration of Lys-sensitive binding of ArgP to lysP was made under the identicalconditions in which ArgP’s binding to argO was Lys insensitive(Fig. 1).
Unique features of argO regulation by ArgP. Of the differentgenes under the control of ArgP in vivo, argO appears to beunique in at least two ways in its manner of regulation. First, itis the only gene that requires Arg as a coeffector for its acti-vation. Second, it is also the only example in which ArgP’sbinding to the cis regulatory region is not Lys sensitive, so that
TABLE 8. Tests for cross-regulation by ArgP and LysR
Strain (genotype)
�-Gal sp act (Miller units)a
lysA-lac with: lysR-lac with: lysP-lacwith NilNil Arg Lys Nil Arg Lys
GJ9650 (wild type) 215 300 68 30 33 33 1,004GJ9651 (�argP) 72 75 29 29 33 34 49GJ9652 (�lysR) 23 26 22 59 65 63 1,052GJ9653 (�argP �lysR) 24 23 25 66 66 65 49
a Values reported are the specific activities of �-galactosidase in derivatives ofthe indicated strains carrying both the Para-lysA plasmid pHYD2673 and one ofthe following plasmids: pHYD2670 (lysA-lac), pHYD2675 (lysR-lac), orpHYD2636 (lysP-lac). Strains were grown in minimal A medium supplementedwith 0.2% each of glycerol and l-arabinose along with one of the following: 18amino acids other than Arg and Lys (Nil), 18 amino acids and 1 mM Arg (Arg),or 18 amino acids and 1 mM Lys (Lys).
TABLE 9. Tests for ArgP regulation of genes relatedto Arg transport
Plasmid (genotype)
�-Gal sp act (Miller units)a
Wild-type strainwith: �argP strain with: argR
with ArgNil Arg Lys Nil Arg Lys
pHYD2658 (artJ-lac) 1,353 286 1,480 1,256 231 1,252 1,700pHYD2659 (artP-lac) 170 105 197 184 122 210 190pHYD2660 (argT-lac) 249 279 242 279 255 221 198pHYD2661 (hisJ-lac) 227 115 262 244 114 296 289
a Values reported are the specific activities of �-galactosidase in derivatives ofMC4100 (wild type), GJ9602 (�argP), and GJ4748 (argR) carrying the indicatedlac fusion plasmid after growth in glucose-minimal A medium without (Nil) orwith supplementation with Arg or Lys at 1 mM.
TABLE 6. ArgP regulation of lac fusions to nested deletions of lysP
Plasmid (lysP extent)a
�-Gal sp act (Miller units)b
argP� strain �argP strain
Nil Lys Nil Lys
pHYD2636 (�219 to �32) 802 150 31 30pHYD2647 (�114 to �32) 852 122 52 50pHYD2648 (�76 to �32) 43 47 41 42
a Extents of lysP shown are in base pairs relative to the start site of transcrip-tion.
b Values reported are the specific activities of �-galactosidase (�-Gal) in de-rivatives of MC4100 (argP�) or GJ9602 (�argP) carrying the indicated lysP-lacfusion plasmids after growth in glucose-minimal A without (Nil) or with (Lys)Lys supplementation at 1 mM.
TABLE 7. Effects of CadC, LysP, and ArgP oncadBA-lac expression
Strain (genotype)
�-Gal sp act (Miller units)a
pH 7.4 pH 5.8
Nil Lys Nil Lys
MC4100 (wild type) 25 28 24 98GJ9623 (�lysP) 27 27 94 96GJ9647 (�cadC) 25 26 26 24GJ9602 (�argP) 23 22 26 96GJ9649 (�lysP �cadC) 21 28 24 26GJ9648 (�argP �cadC) 25 26 24 28GJ9624 (�argP �lysP) 27 24 87 99
a Values reported are the specific activities of �-galactosidase (�-Gal) in theindicated strain derivatives carrying the cadBA-lac fusion plasmid pHYD2674after growth in minimal A-based medium of pH 7.4 or pH 5.8 supplemented witheither 19 amino acids other than Lys (Nil) or 19 amino acids and 10 mM Lys(Lys).
VOL. 193, 2011 ArgP AND LYSINE REPRESSION IN E. COLI 5993
on Novem
ber 10, 2011 by CE
NT
RE
FO
R D
NA
FIN
GE
RP
RIN
TIN
G A
ND
DIA
GN
OS
TIC
S (C
DF
D)
http://jb.asm.org/
Dow
nloaded from
repression by Lys is achieved by an RNA polymerase-trappingmechanism at the argO promoter.
The differences between argO and the other target geneswith regard to the features of their regulation by ArgP are alsoreflected in their differential responses to the argPd mutations(although it must be noted that these mutations were selectedon the basis of their ability to confer greatly increased argOexpression). Perhaps the most prominent distinction is thatobtained with the P274S variant of ArgP, as discussed above.
Thus, it appears that the mechanism of regulation of argO byArgP is fundamentally different from that of the other genesregulated by this protein. At the same time, it is noteworthythat two of the genes that are most prominently regulated byArgP, argO and lysP, are both also regulated by the leucine-responsive general transcriptional regulator Lrp (37, 41).
dnaA and nrdA are not regulated by ArgP in vivo. In itsproposed avatar as IciA, ArgP has been reported to bind oriCin both E. coli (21–23, 46) and Mycobacterium tuberculosis (25)and to regulate transcription from promoters of the dnaA andnrdA genes (20, 27–29). We found that ArgP does bind to theregulatory regions of dnaA and nrdA in vitro with moderateaffinity but that it does not regulate their transcription in vivo.
Thus, we believe that ArgP does not play a role in transactionsinvolving DNA metabolism or replication in E. coli, whichconclusion is supported by the facts that neither a �argP mu-tant nor an ArgP-overproducing strain is compromised forDNA replication or growth rate (46).
ArgP as a noncanonical transcriptional regulator. Genome-wide chromatin immunoprecipitation studies have been de-scribed recently for a number of transcriptional regulators in E.coli, according to which several, such as MelR (17, 18), MntR(49), NsrR (35), and PurR (12), behave canonically in thateach exhibits binding at sites where it serves to control tran-scription of the adjacent genes or operons. Lrp also appears tobe a canonical transcriptional regulator, although it bindsnearly 140 sites in the genome (11).
On the other hand, DNA binding by the activator proteinC-reactive protein (CRP) is apparently noncanonical, sincethere are about 70 sites where it binds strongly and exerts itstranscriptional regulatory function and a further approximately104 sites of low-affinity binding (17). CRP’s behavior contrastswith that of its closely related paralog FNR, which binds ca-nonically at around 65 locations without any noise of low-affinity binding (16). Another example of noncanonical binding
FIG. 3. EMSAs with ArgP and cis regulatory regions of different genes in the absence or presence of the coeffector Lys. ArgP monomerconcentrations are indicated for each lane. Bands corresponding to free DNA and to DNA in binary complex with ArgP (the latter is not observedfor lacZ) are marked by filled and open arrowheads, respectively. Lanes 10 to 12 in panels B to D represent the results of control EMSA reactionswith lysP (full-length) or dapD templates as marked.
5994 MARBANIANG AND GOWRISHANKAR J. BACTERIOL.
on Novem
ber 10, 2011 by CE
NT
RE
FO
R D
NA
FIN
GE
RP
RIN
TIN
G A
ND
DIA
GN
OS
TIC
S (C
DF
D)
http://jb.asm.org/
Dow
nloaded from
is RutR, many of whose binding sites fall within gene codingregions (43).
Our studies would suggest that the ArgP or IciA protein isalso a noncanonical transcriptional regulator, in that it doesexhibit specific DNA binding to mediate the transcriptionalregulation of particular genes in vivo, but in addition, it alsobinds at other sites in the genome that are not associated withregulatory outcomes. The purpose of the latter category ofbinding by ArgP remains to be determined. It has been spec-ulated in the case of CRP that by bending the DNA at itsnoncanonical binding sites, the protein may contribute to chro-mosome shaping and chromatin compaction (17, 47). ArgP toomay similarly participate in shaping the nucleoid architecture,given its reported propensity to bind curved DNA (2). It is alsopossible that the putative dual functions of ArgP, in transcrip-tional regulation and in chromosome organization, are in someway related to its ability to dimerize in two different modes (asdeduced from the crystal structure of the M. tuberculosis or-tholog) (50).
Finally, it is noteworthy that many of the genes regulated byArgP are controlled also by other mechanisms, such as by LysRfor lysA, osmolarity for gdhA, the riboswitch for lysC, and Lrpfor argO and lysP. These results therefore lend support to thenotion that multitarget regulators and multifactor promotersrepresent the norm in bacterial gene regulation (24).
ACKNOWLEDGMENTS
We thank members of the Laboratory of Bacterial Genetics for theiradvice and discussions.
C.N.M. and J.G. are recipients, respectively, of a doctoral ResearchFellowship from the Council of Scientific and Industrial Research andthe J. C. Bose Fellowship from the Department of Science and Tech-nology, Government of India. This work was supported by a Centre ofExcellence in Microbial Biology research grant from the Departmentof Biotechnology, Government of India.
REFERENCES
1. Andrews, A. E., B. Dickson, B. Lawley, C. Cobbett, and A. J. Pittard. 1991.Importance of the position of TYR R boxes for repression and activation ofthe tyrP and aroF genes in Escherichia coli. J. Bacteriol. 173:5079–5085.
2. Azam, T. A., and A. Ishihama. 1999. Twelve species of the nucleoid-associ-ated protein from Escherichia coli. Sequence recognition specificity andDNA binding affinity. J. Biol. Chem. 274:33105–33113.
3. Azam, T. A., A. Iwata, A. Nishimura, S. Ueda, and A. Ishihama. 1999.Growth phase-dependent variation in protein composition of the Escherichiacoli nucleoid. J. Bacteriol. 181:6361–6370.
4. Baba, T., et al. 2006. Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol. Syst. Biol. 2:1–11.
5. Bellmann, A., et al. 2001. Expression control and specificity of the basicamino acid exporter LysE of Corynebacterium glutamicum. Microbiology147:1765–1774.
6. Berlyn, M. K. 1998. Linkage map of Escherichia coli K-12, edition 10: thetraditional map. Microbiol. Mol. Biol. Rev. 62:814–984.
7. Bouvier, J., P. Stragier, V. Morales, E. Remy, and C. Gutierrez. 2008. Lysinerepresses transcription of the Escherichia coli dapB gene by preventing itsactivation by the ArgP activator. J. Bacteriol. 190:5224–5229.
8. Caldara, M., D. Charlier, and R. Cunin. 2006. The arginine regulon ofEscherichia coli: whole-system transcriptome analysis discovers new genesand provides an integrated view of arginine regulation. Microbiology 152:3343–3354.
9. Caldara, M., P. N. Minh, S. Bostoen, J. Massant, and D. Charlier. 2007.ArgR-dependent repression of arginine and histidine transport genes inEscherichia coli K-12. J. Mol. Biol. 373:251–267.
10. Celis, R. T. 1999. Repression and activation of arginine transport genes inEscherichia coli K 12 by the ArgP protein. J. Mol. Biol. 294:1087–1095.
11. Cho, B. K., C. L. Barrett, E. M. Knight, Y. S. Park, and B. O. Palsson. 2008.Genome-scale reconstruction of the Lrp regulatory network in Escherichiacoli. Proc. Natl. Acad. Sci. U. S. A. 105:19462–19467.
12. Cho, B. K., et al. 2011. The PurR regulon in Escherichia coli K-12 MG1655.Nucleic Acids Res. 39:6456–6464.
13. Csonka, L. N., and W. Epstein. 1996. Osmoregulation, p. 1210–1223. In F. C.Neidhardt et al. (ed.), Escherichia coli and Salmonella: cellular and molec-ular biology, 2nd ed. American Society for Microbiology, Washington, DC.
14. Datsenko, K. A., and B. L. Wanner. 2000. One-step inactivation of chromo-somal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad.Sci. U. S. A. 97:6640–6645.
15. Goss, T. J. 2008. The ArgP protein stimulates the Klebsiella pneumoniaegdhA promoter in a lysine-sensitive manner. J. Bacteriol. 190:4351–4359.
16. Grainger, D. C., H. Aiba, D. Hurd, D. F. Browning, and S. J. Busby. 2007.Transcription factor distribution in Escherichia coli: studies with FNR pro-tein. Nucleic Acids Res. 35:269–278.
17. Grainger, D. C., D. Hurd, M. Harrison, J. Holdstock, and S. J. Busby. 2005.Studies of the distribution of Escherichia coli cAMP-receptor protein andRNA polymerase along the E. coli chromosome. Proc. Natl. Acad. Sci.U. S. A. 102:17693–17698.
18. Grainger, D. C., et al. 2004. Genomic studies with Escherichia coli MelRprotein: applications of chromatin immunoprecipitation and microarrays. J.Bacteriol. 186:6938–6943.
19. Guzman, L. M., D. Belin, M. J. Carson, and J. Beckwith. 1995. Tight regu-lation, modulation, and high-level expression by vectors containing the ara-binose PBAD promoter. J. Bacteriol. 177:4121–4130.
20. Han, J. S., H. S. Kwon, J. B. Yim, and D. S. Hwang. 1998. Effect of IciAprotein on the expression of the nrd gene encoding ribonucleoside diphos-phate reductase in E. coli. Mol. Gen. Genet. 259:610–614.
21. Hwang, D. S., and A. Kornberg. 1990. A novel protein binds a key originsequence to block replication of an E. coli minichromosome. Cell 63:325–331.
22. Hwang, D. S., and A. Kornberg. 1992. Opposed actions of regulatory pro-teins, DnaA and IciA, in opening the replication origin of Escherichia coli.J. Biol. Chem. 267:23087–23091.
23. Hwang, D. S., B. Thony, and A. Kornberg. 1992. IciA protein, a specificinhibitor of initiation of Escherichia coli chromosomal replication. J. Biol.Chem. 267:2209–2213.
24. Ishihama, A. 2010. Prokaryotic genome regulation: multifactor promoters,multitarget regulators and hierarchic networks. FEMS Microbiol. Rev. 34:628–645.
25. Kumar, S., A. Farhana, and S. E. Hasnain. 2009. In-vitro helix opening of M.tuberculosis oriC by DnaA occurs at precise location and is inhibited by IciAlike protein. PLoS One 4:e4139.
26. Laishram, R. S., and J. Gowrishankar. 2007. Environmental regulationoperating at the promoter clearance step of bacterial transcription. GenesDev. 21:1258–1272.
27. Lee, Y., H. Lee, J. Yim, and D. Hwang. 1997. The binding of two dimers ofIciA protein to the dnaA promoter 1P element enhances the binding of RNApolymerase to the dnaA promoter 1P. Nucleic Acids Res. 25:3486–3489.
28. Lee, Y. S., and D. S. Hwang. 1997. Occlusion of RNA polymerase by oli-gomerization of DnaA protein over the dnaA promoter of Escherichia coli.J. Biol. Chem. 272:83–88.
29. Lee, Y. S., H. Kim, and D. S. Hwang. 1996. Transcriptional activation of thednaA gene encoding the initiator for oriC replication by IciA protein, aninhibitor of in vitro oriC replication in Escherichia coli. Mol. Microbiol.19:389–396.
30. Maddocks, S. E., and P. C. Oyston. 2008. Structure and function of theLysR-type transcriptional regulator (LTTR) family proteins. Microbiology154:3609–3623.
31. Miller, J. H. 1992. A short course in bacterial genetics: a laboratory manualand handbook for Escherichia coli and related bacteria. Cold Spring HarborLaboratory, Cold Spring Harbor, NY.
32. Nandineni, M. R., and J. Gowrishankar. 2004. Evidence for an arginineexporter encoded by yggA (argO) that is regulated by the LysR-type tran-scriptional regulator ArgP in Escherichia coli. J. Bacteriol. 186:3539–3546.
33. Nandineni, M. R., R. S. Laishram, and J. Gowrishankar. 2004. Osmosensi-tivity associated with insertions in argP (iciA) or glnE in glutamate synthase-deficient mutants of Escherichia coli. J. Bacteriol. 186:6391–6399.
34. Neely, M. N., C. L. Dell, and E. R. Olson. 1994. Roles of LysP and CadC inmediating the lysine requirement for acid induction of the Escherichia colicad operon. J. Bacteriol. 176:3278–3285.
35. Partridge, J. D., D. M. Bodenmiller, M. S. Humphrys, and S. Spiro. 2009.NsrR targets in the Escherichia coli genome: new insights into DNA se-quence requirements for binding and a role for NsrR in the regulation ofmotility. Mol. Microbiol. 73:680–694.
36. Patte, J.-C. 1996. Biosynthesis of threonine and lysine, p. 528–541. In F. C.Neidhardt et al. (ed.), Escherichia coli and Salmonella: cellular and molec-ular biology, 2nd ed. American Society for Microbiology, Washington, DC.
37. Peeters, E., P. N. Le Minh, M. Foulquie-Moreno, and D. Charlier. 2009.Competitive activation of the Escherichia coli argO gene coding for an argi-nine exporter by the transcriptional regulators Lrp and ArgP. Mol. Micro-biol. 74:1513–1526.
38. Reitzer, L. 6 July 2004, posting date. Module 3.6.1.3, Biosynthesis of gluta-mate, aspartate, asparagine, L-alanine, and D-alanine. In R. Curtiss III et al.(ed.), EcoSal Escherichia coli and Salmonella: cellular and molecular biology.ASM Press, Washington, DC. doi:10.1128/ecosal.3.6.1.3.
VOL. 193, 2011 ArgP AND LYSINE REPRESSION IN E. COLI 5995
on Novem
ber 10, 2011 by CE
NT
RE
FO
R D
NA
FIN
GE
RP
RIN
TIN
G A
ND
DIA
GN
OS
TIC
S (C
DF
D)
http://jb.asm.org/
Dow
nloaded from
39. Rodionov, D. A., A. G. Vitreschak, A. A. Mironov, and M. S. Gelfand. 2003.Regulation of lysine biosynthesis and transport genes in bacteria: yet anotherRNA riboswitch? Nucleic Acids Res. 31:6748–6757.
40. Rudd, K. E. 1998. Linkage map of Escherichia coli K-12, edition 10: thephysical map. Microbiol. Mol. Biol. Rev. 62:985–1019.
41. Ruiz, J., I. Haneburger, and K. Jung. 2011. Identification of ArgP and Lrpas transcriptional regulators of lysP, the gene encoding the specific lysinepermease of Escherichia coli. J. Bacteriol. 193:2536–2548.
42. Sambrook, J., and D. W. Russell. 2001. Molecular cloning: a laboratorymanual, 3rd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.
43. Shimada, T., A. Ishihama, S. J. Busby, and D. C. Grainger. 2008. TheEscherichia coli RutR transcription factor binds at targets within genes aswell as intergenic regions. Nucleic Acids Res. 36:3950–3955.
44. Sudarsan, N., J. K. Wickiser, S. Nakamura, M. S. Ebert, and R. R. Breaker.2003. An mRNA structure in bacteria that controls gene expression bybinding lysine. Genes Dev. 17:2688–2697.
45. Tetsch, L., C. Koller, I. Haneburger, and K. Jung. 2008. The membrane-integrated transcriptional activator CadC of Escherichia coli senses lysine
indirectly via the interaction with the lysine permease LysP. Mol. Microbiol.67:570–583.
46. Thony, B., D. S. Hwang, L. Fradkin, and A. Kornberg. 1991. iciA, an Esch-erichia coli gene encoding a specific inhibitor of chromosomal initiation ofreplication in vitro. Proc. Natl. Acad. Sci. U. S. A. 88:4066–4070.
47. Wade, J. T., K. Struhl, S. J. Busby, and D. C. Grainger. 2007. Genomicanalysis of protein-DNA interactions in bacteria: insights into transcriptionand chromosome organization. Mol. Microbiol. 65:21–26.
48. Wei, T., and R. Bernander. 1996. Interaction of the IciA protein with AT-rich regions in plasmid replication origins. Nucleic Acids Res. 24:1865–1872.
49. Yamamoto, K., A. Ishihama, S. J. Busby, and D. C. Grainger. 2011. TheEscherichia coli K-12 MntR miniregulon includes dps, which encodes themajor stationary-phase DNA-binding protein. J. Bacteriol. 193:1477–1480.
50. Zhou, X., et al. 2010. Crystal structure of ArgP from Mycobacterium tuber-culosis confirms two distinct conformations of full-length LysR transcrip-tional regulators and reveals its function in DNA binding and transcriptionalregulation. J. Mol. Biol. 396:1012–1024.
5996 MARBANIANG AND GOWRISHANKAR J. BACTERIOL.
on Novem
ber 10, 2011 by CE
NT
RE
FO
R D
NA
FIN
GE
RP
RIN
TIN
G A
ND
DIA
GN
OS
TIC
S (C
DF
D)
http://jb.asm.org/
Dow
nloaded from